JPH09214767A - Picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve gradation reproductivity and, moreover to visibly reduce noise existing in an input picture itself regardless of a picture output device for outputting an output picture signal where noise is superimposed and also regardleses of the contents of an input picture and the kind of a picture input device when noise which is hard to be visually sensed is positively superimposed on a picture signal. SOLUTION: Noise data corresponding to the picture output devices 5A and 5B is previously written in an output device correspondence noise providing part 80. An input picture signal feature quantity calculating part 60 calculates the noise feature quantities of a lightness signal L and chromaticity signals (a) and (b). A noise quantity calculating part 70 calculates the noise quantity from the calculated feature quantities dL, da and db,; noise N1 corresponding to the picture output device outputting the output picture signal So from the output device correspondence noise providing part 80, the lightness signal L and the chromaticity signals (a) and (b). The calculated noise quantities N2L, N2a and N2b are superimposed with the lightness signal L and the chromaticity signals (a) and (b) in a noise superimposing part 50.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an input image signal when converting an input image signal from an image input device such as an image scanner or a video camera into an output image signal output by an image output device such as a printer or a display. The present invention relates to an image processing device that superimposes noise on.

[0002]

2. Description of the Related Art Color reproducibility, gradation reproducibility, sharpness and graininess are important characteristics in considering the quality of an image output by a color copying machine. Many image processing techniques for faithfully reproducing an input image such as an image on a document with an output image have been considered.

If an image is used as an information transmission means,
The final receiving system is the human visual system. Therefore, it is desirable that the above image processing considers the perceptual characteristics based on the human visual psychological phenomenon.

Therefore, as for color reproduction, for example, as shown in Japanese Patent Laid-Open No. 3-234178, L ^* a close to the human visual sense so that more faithful and preferable color reproduction characteristics can be obtained. It is considered to perform color reproduction correction in the ^* b ^* space.

The L ^* a ^* b ^* space, its lightness signal L ^* , and its chromaticity signals a ^* and b ^* are indicated by omitting * from now on.

Regarding the sharpness, even if the high frequency component which does not respond to the visual sense is removed from the spatial frequency response characteristic of the visual sense, the image quality is not affected, and it is efficient to do so.
(Modulation transfer function) It is said to be a design method.

Regarding the graininess, from the same idea as the sharpness, it is generally performed in the sharpness enhancement filter by cutting the high frequency band to make the sharpness and the graininess compatible with each other. .

Further, regarding the relationship between the image noise and the gradation, it is known from the human visual characteristics that the presence of the image noise has the effect of blocking the gradation step, so-called pseudo contour. This is applied in the form of the dither method in the halftone generation.

Recently, for example, Japanese Patent Laid-Open No. 4-129
As disclosed in Japanese Patent No. 384 and Japanese Patent Laid-Open No. 5-114999, noise that is hard to visually perceive is positively superimposed on the image signal in consideration of the noise perceptual characteristics of the visual system. An image processing apparatus has been considered which can improve gradation reproducibility without impairing color tone and sharpness, and can also visually reduce noise existing in the input image itself.

[0010]

As an image system,
As shown in FIG. 1, an image input device 1, an image processing device 3, and a plurality of image output devices 5A and 5B are connected via a network 7, and an input image signal S from the image input device 1 is input.
The image processing device 3 converts i into an output image signal So output by the image output device 5A or 5B, and the output image signal So is selected by the image output device 5 selected by the user.
What is output by A or 5B is considered.

In such an image system, when noise is superimposed on the input image signal Si in the image processing apparatus 3 as described above, the image output apparatuses 5A and 5B have the same gradation reproduction characteristics, Since the respective output noise characteristics may be different, for example, when the image output device 5A is easy to amplify the noise, when the image output device 5A outputs the output image signal So on which noise is superimposed,
When the amount of superimposed noise is relatively large, the apparent image quality is returned and deteriorated, and conversely, when the image output device 5A is difficult to amplify noise, the output image signal So is output by the image output device 5A. When doing so, the amount of noise superposition becomes relatively small, and the image quality improvement due to the superposition of noise becomes insufficient.

In addition, noise existing in the input image itself,
The influence on the image quality due to the superposition of noise on the input image signal Si differs depending on the contents of the input image and the type of the image input device 1, and depending on the contents of the input image and the type of the image input device 1, the noise may still occur. Superimposing the image on the screen deteriorates the apparent image quality, or the image quality is not sufficiently improved by superimposing noise.

Therefore, according to the present invention, in consideration of the perceptual characteristics based on the human visual psychology phenomenon, when noise that is difficult to visually perceive is positively superimposed on the image signal, the output image on which the noise is superimposed is Regardless of the image output device that outputs the signal, and regardless of the content of the input image or the type of the image input device, the tone reproducibility is improved without impairing the color tone and sharpness. The noise existing in the input image itself can be visually reduced.

[0014]

In the present invention, referring to the reference numerals of the embodiment shown in FIG. 2 and described later, the output for outputting the input image signal Si from the image input device 1 by the image output device 5A or 5B. In the image processing device 3 for converting into the image signal So, the input image signal feature amount calculating means 60 for calculating the grayscale distribution feature amount of the input image signal Si and the noise N1 corresponding to the image output device 5A or 5B are provided. Output device-corresponding noise providing unit 80, the grayscale distribution feature amount calculated by the input image signal feature amount calculating unit 60, the output device corresponding noise N1 provided by the output device corresponding noise providing unit 80, and the input image signal Noise amount calculating means 70 for calculating the amount of noise superimposed on the input image signal Si based on Si, and this noise amount calculating means Noise N2 of the calculated noise amount by 0 a noise superimposing means 50 to be superimposed on the input image signal Si, the provision.

[0015]

In the image processing apparatus 3 of the present invention configured as described above, the noise amount calculating means 70 outputs the output image signal So by the output device corresponding noise N1 provided by the output device corresponding noise providing means 80. The noise superposition amount is determined according to the output noise characteristic of the image output device 5A or 5B to be output.

Therefore, for example, the image output device 5A is a device which easily amplifies high frequency noise.
When the output image signal So is output at A, the superposition amount of high frequency noise on the input image signal Si is suppressed, and conversely, the image output device 5A hardly amplifies the high frequency noise. When outputting the output image signal So, the output image signal S on which the noise N2 is superimposed, such as the amount of high frequency noise superimposed on the input image signal Si is enhanced.
Regardless of the output noise characteristic of the image output apparatus that outputs o, image quality deterioration due to excessive noise superimposition amount and insufficient image quality improvement due to insufficient noise superimposition amount do not occur.

At the same time, the noise amount calculating means 70 determines the noise superimposing amount according to the density distribution characteristic amount of the input image signal Si calculated by the input image signal characteristic amount calculating means 60. In areas of characters and line drawings where image quality is likely to deteriorate due to image noise superposition, or in areas of uniform density, the amount of noise superposition is suppressed, and gradation steps in the halftone of the input image, so-called pseudo contour areas Then, by increasing the noise superposition amount, etc., for any part of the grayscale distribution of the input image,
Gradation reproducibility is improved without impairing color tone and sharpness,
Furthermore, the noise present in the input image itself is also visually reduced.

[0018]

FIG. 2 shows an example of a network image system using an example of the image processing apparatus of the present invention. The image system is the same as that shown in FIG.
The image input device 1, the image processing device 9 and the plurality of image output devices 5A and 5B are connected via a network 7, and the input image signal Si from the image input device 1 is transferred by the image processing device 9 to the image output device 5A or The output image signal So is output by 5B, and the output image signal So is output by the image output device 5A or 5B selected by the user.

The image processing apparatus 9 as an example of the image processing apparatus of the present invention includes a control unit 10 and a noise superimposing unit 5.
0, input image signal feature amount calculation unit 60, noise amount calculation unit 7
0 and the noise providing unit 80 corresponding to the output device, and the input image signal S from the image input device 1 through the network 7.
i is taken into the image processing apparatus 9 through the control unit 10, and the noise superimposing unit 50 and the input image signal feature amount calculating unit 6 are acquired.
0 and the noise amount calculation unit 70, and the input image signal characteristic amount calculation unit 60 calculates the grayscale distribution characteristic amount of the input image signal Si, and the calculated characteristic amount is calculated with the input image signal Si. Is supplied to the unit 70.

The output device-corresponding noise providing unit 80 includes output device-corresponding noise storage units 81A and 81B and an output device-corresponding noise reading unit 82, as shown in FIG. , 81B are respectively written in advance with noise data corresponding to the image output devices 5A and 5B, which are generated based on the noise transfer characteristics of the image output devices 5A and 5B.

Then, a signal SE for selecting one of the image output devices 5A and 5B is supplied from the control unit 10 to the output device-corresponding noise reading unit 82 according to a user's instruction.
The output device-compatible noise reading unit 82 outputs the output image signal So from either the output device-compatible noise storage unit 81A or 81B, that is, the image processing device 9 to the image output device 5.
When outputting at A, the output device corresponding noise storage unit 81A
Therefore, when the output image signal So is output by the image output device 5B, the noise data is read from the output device corresponding noise storage section 81B, and the read noise data is the noise amount as the output device corresponding noise N1. Calculation unit 70
Is supplied to.

In the noise amount calculation unit 70, the density distribution characteristic amount from the input image signal characteristic amount calculation unit 60, the output device corresponding noise N1 from the output device corresponding noise providing unit 80, and the input image signal Si from the control unit 10 are inputted. Based on this, the noise amount superimposed on the input image signal Si is calculated, and the noise amount N2 of the noise amount calculated by the noise amount calculation unit 70 is calculated.
Is superposed on the input image signal Si in the noise superimposing unit 50, and the output image signal So on which the noise N2 is superposed
Is sent to the image output device 5A or 5B selected by the user through the control unit 10 and the network 7, and is output.

As will be described in more detail below, in the noise amount calculation section 70, when the output device corresponding noise N1 from the output device corresponding noise providing portion 80 is large, the input image signal S
When the amount of noise superimposed on i is small and the noise N1 corresponding to the output device is small, the amount of noise superimposed on the input image signal Si is increased, and the grayscale distribution characteristic amount from the input image signal characteristic amount calculation unit 60 is increased. Indicates a character or line drawing area of the input image or an area with uniform density,
The amount of noise superimposed on the input image signal Si is reduced,
When the grayscale distribution feature amount indicates the halftone region of the input image, the amount of noise superimposed on the input image signal Si is increased.

In the case of a color image, the perceptual characteristics based on the human visual psychological phenomenon with respect to noise differ depending on each color component or according to hue, saturation, or brightness. Therefore, in the case of a color image, for example, a small amount of noise is superimposed on a color component having a high perception of noise, and a large amount of noise is superimposed on a color component having a low perception of noise. , Or that the noise is superimposed only on the color components that are less perceptible to noise, it is desirable to control the amount of superimposed noise for each color component or according to hue, saturation, or brightness. .

FIG. 3 shows an example of an image processing apparatus in the case of a color image in consideration of this, and the input image signal Si
Is a color image signal composed of red, green, and blue image signals R, G, and B, and the input image signal Si is supplied to the lightness / chromaticity separation color conversion unit 30, and the lightness signal L in the Lab space described above is supplied. And the chromaticity signals a and b are separated and converted, and the lightness signal L and the chromaticity signals a and b are supplied to the input image signal characteristic amount calculation unit 60, and in the input image signal characteristic amount calculation unit 60, Is the lightness signal L, as will be described later.
And the grayscale distribution feature amount of the input image signal Si is calculated for each of the chromaticity signals a and b.

In the noise amount calculation unit 70, which will be described later in detail, in the brightness noise amount calculation unit 70L, the density distribution characteristic amount dL for the lightness signal L from the input image signal characteristic amount calculation unit 60, Based on the output device corresponding noise N1 from the output device corresponding noise providing unit 80 and the lightness signal L from the lightness / chromaticity separation / color conversion unit 30, the noise amount to be superimposed on the lightness signal L is calculated, and the chromaticity is calculated. In the noise amount calculation unit 70ab, the grayscale distribution feature amount da for the chromaticity signals a and b from the input image signal feature amount calculation unit 60,
db, the output device-corresponding noise N1 from the output device-corresponding noise providing unit 80, and the chromaticity signals a and b from the lightness / chromaticity separation / color converting unit 30 based on the amount of noise superimposed on the chromaticity signals a and b. Is calculated.

In the noise superimposing unit 50, the noise amount noise N2L calculated by the lightness noise amount calculating unit 70L of the noise amount calculating unit 70 is superimposed on the lightness signal L from the lightness / chromaticity separation / color conversion unit 30. At the same time, the noise amount noises N2a and N2b calculated by the chromaticity noise amount calculation unit 70ab are calculated.
Are superposed on the chromaticity signals a and b from the brightness / chromaticity separation color conversion unit 30, respectively, and the noise signal N2L and N2a and N2b from the noise superposition unit 50 are superposed on the brightness signal L.
n and the chromaticity signals an and bn are sent to the image output device as the output image signal So.

FIG. 4 shows a practical example of a color image system including a more practical example of such an image processing apparatus for color images.

The image input device 1 is a device such as a color image scanner or a color video camera which optically reads a document or an object and converts it into a color image signal. The image reading unit 101 and each unit of the device are controlled to read an image. The color image signal from the unit 101 is sent to the image processing apparatus 9 via the network 7 as the input image signal Si.

Input image signal Si from the image input device 1
Is red quantized to 8 bits for each pixel,
It is composed of green and blue image signals R, G and B.

The image output devices 5A and 5B respectively control the respective parts of the device, and as will be described later, the control parts 501A and 501B from which the output image signal So sent from the image processing device 9 through the network 7 is taken into the device. Subtractive color conversion unit 502A for performing subtractive color conversion of the image signal of
502B, undercolor removing units 503A and 503B that remove the undercolor of the image signal from now on, halftone generating units 504A and 504B that generate halftone for the image signal of the future, and an image output unit that prints out the image signal of the future. 505
This is a general image output device including A and 505B.

Then, the image processing device 9 includes a control unit 10,
The gradation conversion unit 20, the lightness / chromaticity separation color conversion unit 30, the sharpness correction unit 40, the noise superposition unit 50, the input image signal feature amount calculation unit 60, the noise amount calculation unit 70, and the output device compatible noise providing unit 80 are provided. The control unit 10 is composed of, for example, a computer system such as a personal computer having a user interface and a control program, and has the functions as described above with reference to FIG.

The gradation conversion unit 20 uses the conversion table to
The image signals R, G, B from the control unit 10 are converted to the equivalent brightness signal R
e, Ge, Be gradation conversion, lightness chromaticity separation color conversion unit 3
0 is the matrix calculation or conversion table, and the equivalent lightness signals Re, Ge, Be from the gradation conversion unit 20 are separated and converted into the lightness signal L and the chromaticity signals a, b in the Lab space as described above with reference to FIG. To do.

The sharpness correction unit 40 performs the convolution of the image and the unsharpness mask to obtain the lightness signal L and the chromaticity signal a from the lightness / chromaticity separation color conversion unit 30.
The sharpness of b is corrected, and in this example, the brightness signal L
For example, the sharpness is emphasized by an unsharpness mask filter having a spatial frequency characteristic as shown in FIG.
As for each of the above items, the sharpness of each of them is reduced by an unsharpness mask filter having a spatial frequency characteristic as shown in FIG. 11B.

In this example, the input image signal characteristic amount calculation unit 60 changes the lightness signal L and the chromaticity signals a and b from the lightness / chromaticity separation color conversion unit 30 by the first-order differential operation by the Sobel operator. The amount is calculated, and the calculated amount of change is set as the characteristic amount for the lightness signal L and the chromaticity signals a and b.

FIG. 5 shows an example of such an input image signal characteristic amount calculating section 60. The input image signal characteristic amount calculating section 60 includes a lightness signal characteristic amount calculating section 60L and a chromaticity signal characteristic amount calculating section 60ab. And the brightness signal feature amount calculation unit 6
At 0L, the lightness signal L from the lightness / chromaticity separation color conversion unit 30 is written in the line buffer 61L, the lightness signal from the line buffer 61L is smoothed by the smoothing circuit 62L, and the lightness signal from the smoothing circuit 62L. But Figure 6
A vertical Sobel operator 63L having a filtering table as shown in FIG. 6A and a horizontal Sobel operator 64L having a filtering table as shown in FIG. Absolute value of vertical differential value from vertical Sobel operator 63L and horizontal Sobel operator 64L
The absolute value of the differential value in the lateral direction from the absolute value addition circuit 65L
And the change amount dL that is the addition output is
It is set as a feature amount for the lightness signal L.

In the chromaticity signal characteristic amount calculating section 60ab, the chromaticity signals a and b from the lightness / chromaticity separation color converting section 30 are calculated in the same manner as the variation amounts da and db as characteristic amounts. It

As shown in FIG. 2 and as shown in FIG. 7, the output device corresponding noise providing unit 80 includes output device corresponding noise storage units 81A and 81B, and an output device corresponding noise reading unit 82. The output device corresponding noise storage section 8
Noise data corresponding to the image output devices 5A and 5B are written in the 1A and 81B in advance. Then, as described above, depending on which of the image output devices 5A and 5B outputs the output image signal So from the image processing device 9, the output device corresponding noise reading section 82 causes the output device corresponding noise storage section 81A or The noise data is read from 81B as the output device corresponding noise N1.

Here, the output device-corresponding noise storage section 81
The noise data stored in A and 81B are respectively standardized to the standard deviation σ = 1, and are generated based on the noise transfer characteristics of the image output devices 5A and 5B as described above.

For example, the image output devices 5A, 5
At B, an output image signal obtained by superimposing random noise σ = 5 on an image signal having average brightness L = 50, a = 0, b = 0 is output, and the noise transfer characteristics of the image output devices 5A and 5B are measured. The analysis is performed, and the spatial frequency characteristic of the superposition noise data to be written in the output device corresponding noise storage units 81A and 81B is determined from the result of the analysis.

FIGS. 12A and 12B show the noise transfer characteristics of the image output devices 5A and 5B thus obtained (however, the screen spectrum is excluded),
The image output device 5A having a noise transfer characteristic as shown in FIG. 12 (A) is an image output device based on a general xerographic process, and it can be seen that it is a device that easily amplifies low and medium frequency noise. The image output device 5B as shown in FIG. 12B is an image output device by a photographic process, and it can be seen that noise is not amplified in any band other than the extremely low band.

Therefore, in this case, the image output device 5A
Corresponding to the output device corresponding noise storage section 81A
As for the spatial frequency characteristic of the noise data written in, for example, as shown in FIG. 13 (A), it is assumed that the low-to-mid range up to 4 lp / mm is cut, and the noise storage section corresponding to the output device 5B corresponds to the image output device 5B. As for the spatial frequency characteristic of the noise data written in 81B, for example, as shown in FIG. 13B, it is assumed that the low frequency band up to 2 lp / mm is cut.

FIG. 8 shows an example of the noise amount calculation unit 70. The noise amount calculation unit 70 has the brightness noise amount calculation unit 70L and the chromaticity noise amount calculation unit 70 as shown in FIG.
The brightness noise amount calculation unit 70L includes a memory 71L and a brightness noise amount calculation circuit 74L, and the chromaticity noise amount calculation unit 70ab includes the memory 71L.
a, 71b and a chromaticity noise amount calculation circuit 74ab.

Then, in the memory 71L, a noise amplification factor table 72L in which a noise amplification factor as shown in FIG. 14A is described with respect to the value of the lightness signal L, and the variation amount of the lightness signal L described above. A noise amplification factor table 73L in which a noise amplification factor as shown in FIG. 14B is described with respect to the relative value of dL is stored, and the noise amplification factor is calculated by the lightness signal Lxy from the lightness / chromaticity separation color conversion unit 30. The noise amplification factor for the lightness signal Lxy is read from the table 72L, and the noise amplification factor for the variation amount dLxy is read from the noise amplification factor table 73L according to the variation amount dLxy from the input image signal feature amount calculation unit 60.

In the following, the lightness signal Lxy, its variation dLxy, etc. will be referred to as coordinates (x,
The lightness signal L, the amount of change dL, and the like for the pixel of y).

Then, in the brightness noise amount calculation circuit 74L, from the noise amplification rate from the noise amplification rate tables 72L and 73L and the noise N1xy from the output device corresponding noise providing section 80, NLxy = TLxy × TdLxy × N1xy ( The noise amount NLxy superimposed on the lightness signal Lxy is calculated by the calculation represented by 1).

However, TLxy is the noise amplification factor for the lightness signal Lxy from the noise amplification factor table 72L, and TdLxy is the noise amplification factor for the change amount dLxy of the lightness signal Lxy from the noise amplification factor table 73L.

Memory 71 of chromaticity noise amount calculation unit 70ab
In a, a noise amplification factor table 72a in which a noise amplification factor as shown in FIG. 15A is described with respect to the value of the chromaticity signal a, and a relative value of the above-described change amount da of the chromaticity signal a. On the other hand, for example, the noise amplification factor table 73a in which the noise amplification factor is described as shown in FIG. 15B is stored, and the chromaticity signal axy from the lightness / chromaticity separation color conversion unit 30 is stored.
From the noise amplification factor table 72a, the chromaticity signal a
The noise amplification factor for xy is read, and the variation amount daxy from the noise amplification factor table 73a is calculated according to the variation amount daxy from the input image signal feature amount calculation unit 60.
The noise amplification factor for is read.

Memory 71 of chromaticity noise amount calculation unit 70ab
In b, a noise amplification factor table 72b in which a noise amplification factor as shown in FIG. 16 (A) is described with respect to the value of the chromaticity signal b, and a relative value of the above-described variation amount db of the chromaticity signal b. On the other hand, for example, the noise amplification factor table 73b in which the noise amplification factor is described as shown in FIG. 16B is stored, and the chromaticity signal bxy from the brightness / chromaticity separation color conversion unit 30 is stored.
From the noise amplification factor table 72b, the chromaticity signal b
The noise amplification factor for xy is read, and the variation amount dbxy from the noise amplification factor table 73b is calculated by the variation amount dbxy from the input image signal feature amount calculation unit 60.
The noise amplification factor for is read.

Then, in the chromaticity noise amount calculation circuit 74ab, the noise amplification factor tables 72a, 73a and 72a.
From the noise amplification factors from b and 73b and the noise N1xy from the output device corresponding noise providing unit 80, first, Naxy = Taxy × Tdaxy × N1xy (2) Nbxy = Tbxy × Tdbxy × N1xy (3) The noise amounts Naxy and Nbxy for the chromaticity signals axy and bxy are calculated by the calculation.

Here, Taxy is the noise amplification factor for the chromaticity signal axy from the noise amplification factor table 72a, and Tdaxy is the noise amplification factor for the change amount daxy of the chromaticity signal axy from the noise amplification factor table 73a, Tbxy. Is the noise amplification factor for the chromaticity signal bxy from the noise amplification factor table 72b, Tdbxy
Is the chromaticity signal bx from the noise amplification factor table 73b.
It is a noise amplification factor with respect to a change amount dbxy of y.

Further, in the chromaticity noise amount calculation circuit 74ab, NCxy = {(Naxy) ** 2+ (Nbxy) ** 2} ** (1/2) (4) (4) "** 2" is the square,
"** (1/2)" is the power of (1/2)), the noise amount NCxy with respect to the saturation axis is calculated, and this, the results of equations (2) and (3), and the lightness color From the chromaticity signals axy and bxy from the degree separation color conversion unit 30, Ncaxy = sign (N1xy) × NCxy × | axy | / {(axy) ** 2+ (bxy) ** 2} ** (1/2 ) (5) Ncbxy = sign (N1xy) × NCxy × | bxy | / {(axy) ** 2+ (bxy) ** 2} ** (1/2) ... (6) Noise amounts Ncaxy and Ncbxy only in the saturation axis direction with respect to the chromaticity signals axy and bxy
Is calculated. However, sign (N1xy) indicates whether the noise data N1xy from the output device corresponding noise providing unit 80 is positive or negative.

In the noise superimposing section 50, as shown in FIG. 9, only the noise amount NLxy for the lightness signal Lxy and the saturation axis direction for the chromaticity signals axy, bxy calculated by the noise amount calculating section 70 as described above. Noise amount Nc
axy, Ncbxy is nLxy = Lxy + NLxy (7) naxy = axy + Ncaxy (8) nbxy = bxy + Ncbxy (9) As shown in (9), the adding circuits 51L and 51a, 51a
In b, the lightness signal Lxy and the chromaticity signals axy and bxy are added to the lightness signal Lxy and the chromaticity signals axy and bxy from the lightness / chromaticity separation / color conversion unit 30 to obtain a lightness signal nLxy and chromaticity signals anxy and bnxy on which noises are respectively superimposed.

As described above, in the color image system shown in FIG. 4, the input image signal Si from the image input device 1 is converted into the lightness signal L and the chromaticity signal a by the lightness / chromaticity separation color conversion unit 30 of the image processing device 9. The input image signal feature amount calculation unit 60, the output device corresponding noise providing unit 80, the noise amount calculation unit 70, and the noise superimposing unit 50 of the image processing device 9 are separated and converted into b.
However, as described above with reference to FIG. 5 and above, the configuration corresponding to the lightness signal L and the chromaticity signals a and b allows the superposition of noise on the input image signal to be a perceptual characteristic based on a human visual psychological phenomenon. Is controlled in a more suitable state,
A more preferable output image can be visually obtained.

However, in the examples shown in FIGS. 8 and 9, the noise is superimposed on the lightness axis and the saturation axis separately, and is expressed by the above equations (1), (2) and (3). Noise amount N
Lxy, Naxy, Nbxy are used as they are for the lightness signal Lx.
y, Lab so as to be superimposed on the chromaticity signals axy and bxy
Noise may be superimposed on the L axis, the a axis, and the b axis in space completely independently. It is also possible to superimpose noise complementarily between the lightness signal and the chromaticity signal.

In the output device corresponding noise providing unit 80 of the examples shown in FIGS. 2 and 7, the noise data corresponding to the image output devices 5A and 5B is directly output device corresponding noise storage unit 81.
A, 81B are stored in the noise storage unit 81 corresponding to the output device.
In the case of reading from A and 81B, the output device corresponding noise providing unit 80 may generate noise data corresponding to the image output devices 5A and 5B based on random noise.

FIG. 10 shows an example of the output device corresponding noise providing unit 80 in this case. The output device corresponding noise providing unit 80 includes a random noise storage unit 91, output device corresponding noise characteristic amount storage units 92A and 92B. An output device-corresponding noise feature amount reading unit 93 and an output device-corresponding noise generating unit 94 are provided, and the output device-corresponding noise feature amount storage units 92A and 92B are provided.
The image output device 5A and the image output device 5A which are calculated from the noise amplification characteristics of the image output devices 5A and 5B when gray random noise is output by superimposing three colors or four colors, respectively. , 5B of the output noise characteristics are written in advance.

Then, according to a user's instruction, a signal SE for selecting one of the image output devices 5A and 5B
The output device-corresponding noise feature amount reading unit 93 reads the above-described feature amount from either of the output device-corresponding noise feature amount storage units 92A and 92B, and the output device-corresponding noise generation unit reads the read feature amount. By converting the spatial characteristics of the random noise from the random noise storage unit 91 according to the amount, the image output device 5A or 5B
Is generated, and the generated noise data is used as the noise N1 from the output device corresponding noise providing unit 80.

[0059]

As described above, according to the present invention, regardless of the image output device that outputs the output image signal on which noise is superimposed, regardless of the input image or the type of the image input device. In addition, it is possible to improve the gradation reproducibility without impairing the color tone and sharpness, and also to visually reduce noise existing in the input image itself.

[Brief description of drawings]

FIG. 1 is a connection diagram showing an example of an image processing system to which the present invention is applied.

FIG. 2 is a block diagram showing an example of an image processing apparatus of the present invention.

FIG. 3 is a block diagram showing an example of the image processing apparatus of the present invention when the input image is a color image.

FIG. 4 is a block diagram showing a more specific example of an image processing system including the image processing device of the present invention when the input image is a color image.

5 is a block diagram showing a specific example of an input image signal characteristic amount calculation unit in FIG.

FIG. 6 is a diagram showing an example of a Sobel operator filtering table used in the input image signal feature amount calculation unit of FIG.

FIG. 7 is a block diagram showing an example of a noise providing unit corresponding to the output device of FIG.

8 is a block diagram showing a specific example of a noise amount calculation unit in FIG.

9 is a block diagram illustrating an example of a noise superimposing unit in FIG.

FIG. 10 is a block diagram showing another example of a noise providing unit corresponding to an output device.

FIG. 11 is a characteristic diagram for explaining the sharpness correction unit in FIG.

FIG. 12 is a characteristic diagram showing an example of noise transfer characteristics of the image output device.

FIG. 13 is a characteristic diagram showing an example of spatial frequency characteristics of noise data corresponding to an image output device.

FIG. 14 is a characteristic diagram showing an example of characteristics of a noise amplification factor with respect to a lightness signal and a change amount thereof.

FIG. 15 is a characteristic diagram showing an example of characteristics of a noise amplification factor with respect to a chromaticity signal and its change amount.

FIG. 16 is a characteristic diagram showing an example of characteristics of a noise amplification factor with respect to a chromaticity signal and its change amount.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input device 5A, 5B Image output device 9 Image processing device 50 Noise superimposing unit 60 Input image signal feature amount calculating unit 70 Noise amount calculating unit 80 Noise providing unit for output device

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hideki Tachibana 430 Nakai-cho, Ashigaragami-gun, Kanagawa Prefecture Green Tech Nakai Fuji Xerox Co., Ltd. Xerox Co., Ltd.

Claims (1)

[Claims] 1. An image processing device for converting an input image signal from an image input device into an output image signal output by an image output device, wherein an input image signal feature amount calculating means for calculating a grayscale distribution feature amount of the input image signal. An output device corresponding noise providing unit that provides noise corresponding to the image output device, a grayscale distribution feature amount calculated by the input image signal feature amount calculating unit, and an output provided by the output device corresponding noise providing unit Noise amount calculation means for calculating the noise amount to be superimposed on the input image signal based on the device-corresponding noise and the input image signal, and the noise amount noise calculated by the noise amount calculation means for the input image signal. An image processing apparatus comprising: a noise superimposing unit that superimposes the noise on the image.